"Meme and Computer Vision" is a Master's thesis work that focuses on the analysis of internet memes using computer vision systems. The study examines six automatic content recognition services, including Amazon Rekognition, Clarifai, Google Vision, Imagga, Keras EfficientNetB7 and Microsoft Azure, in the identification and classification of images. The research analyses how these computer vision systems perceive and classify the various elements within Internet memes, comparing their performance and revealing differences in their interpretation of these multimodal digital artefacts. Using object detection techniques, the study classifies meme elements, extracting information and labels to compare results between different computer vision systems. Furthermore, the thesis explores the ability of these systems to understand the cultural and semiotic context of memes, assessing whether their priorities focus on formal aspects or on the meanings of the content within the context.
